Question 1

What AI models can run on MI350X?

Accepted Answer

The MI350X (288 GB VRAM) can run models including Llama-2-13b, MediPhi-MedWiki, Llama-2-7b, Llama-2-70b, intent-aware-lfqa-qwen3-4b-intent-implicit and more — 10 compatible models indexed so far. Compatible models are determined by whether the model's memory footprint fits within the GPU's 288 GB VRAM, potentially using quantization (Q8, Q6, or Q4) to reduce memory usage.

Question 2

What is the cheapest cloud option for MI350X?

Accepted Answer

The cheapest on-demand option for the MI350X is RunPod at $5.49/hr (1x_MI350X_SECURE). We track pricing across 14 providers to help you find the best deal.

Question 3

Is MI350X good for fine-tuning?

Accepted Answer

Yes, the MI350X is excellent for fine-tuning. With 288 GB of VRAM, it can handle full fine-tuning of models up to ~30B parameters and LoRA/QLoRA fine-tuning of much larger models. It's a top-tier choice for training workloads.

Question 4

How much VRAM does MI350X have?

Accepted Answer

The MI350X has 288 GB of VRAM (HBM3e). It also offers 8000 GB/s memory bandwidth, 2,500 TFLOPS of FP16 compute, FP8 support for efficient inference.

Question 5

What quantization formats work on MI350X?

Accepted Answer

The MI350X supports the following quantization formats: FP16 (full precision), FP8, Q8 (8-bit), Q6 (6-bit), Q4 (4-bit). FP16 works for models up to ~40B parameters. FP8 halves memory vs FP16 with minimal quality loss. Q4 quantization lets models up to ~979B parameters (by VRAM estimate) fit on the MI350X.

MI350X

Specifications

Cloud pricing

Compatible models

Compare this GPU

Frequently asked questions

VRAM	288 GB
Memory type	HBM3e
Bandwidth	8000 GB/s
FP16 TFLOPS	2,500
FP8 support	Yes